Z.ai · 320B (18B active) · Mixture of Experts
First natively multimodal GLM-5 — hybrid sparse/linear attention, 18B active, beats GLM-5.2 at a tenth of the price Check if your GPU or Mac can run GLM-5.3 Flash locally — 178.8 GB min, 298 GB recommended.
2026-081024K context
Mixture of Experts
Total experts: 288
Active experts: 8
Active params: 18.0B
| Quant | Bits | VRAM | Quality | Status |
|---|---|---|---|---|
| Q2_K | 2 | 102.9 GB | low | — |
| Q3_K_M | 3 | 143.9 GB | moderate | — |
| Q4_K_M | 4 | 164.4 GB | good | — |
| Q5_K_M | 5 | 205.4 GB | good | — |
| Q6_K | 6 | 246.4 GB | excellent | — |
| Q8_0 | 8 | 328.3 GB | excellent | — |
| F16 | 16 | 656.2 GB | lossless | — |
Can I run GLM-5.3 Flash locally?
- Can I run GLM-5.3 Flash locally?
- GLM-5.3 Flash needs about 178.8 GB of memory at a minimum and 298 GB recommended. Open this page to grade it against your GPU or Mac, then run it with runai, Ollama or LM Studio.
- How much VRAM does GLM-5.3 Flash need?
- At Q4_K_M, GLM-5.3 Flash uses about 164.4 GB of VRAM. Higher quants need more memory; lower quants fit tighter cards with a quality tradeoff.